This study aims to develop a new index to evaluate the low-temperature cracking performance (LTCP) of recycled asphalt mixtures. To achieve this, the assumption that fracture energy has limitation in evaluating the LTCP of mixtures was put forward firstly. To reveal this limitation theoretically, a concept of energy absorption rate that characterized the fracture energy of the mixtures was then given. Thirdly, an equivalent fracture temperature (EFT) corresponding to the critical cracking temperature in a thermal stress restrained specimen test (TSRST) was proposed to evaluate the LTCP of mixtures based on a three-point bending beam (3PBB) test. Finally, some data derived from previous work were collected to verify the proposed assumption, and the 3PBB test and TSRST were conducted to investigate the LTCP of five recycled mixtures with different reclaimed asphalt pavement (RAP) material contents and to verify the rationality of the proposed EFT. The results confirmed that fracture energy had some limitations in predicting the LTCP of mixtures. The EFT was as accurate as the critical cracking temperature in evaluating the LTCP of mixtures. Compared to fracture energy and critical cracking temperature, EFT had bigger absolute change and relative change between different mixtures, meaning it was better to differentiate the LTCP of mixtures. Compared to failure strain, EFT was also more definite and less variable when processed by different data analysts, meaning it was more objective.